Numerical Optimization of Flight Trajectory for Rockets via Artificial Neural Networks

This research arise to optimize the flight trajectory for rockets, for this were applied hybrid techniques, based on the Finite Difference Method (FDM) to obtain the solution of the non-linear differential equations provided by the analytic modeling. So aiming at the optimizations were applied the Artificial Neural Networks (ANN) into two curves of thrust rocket engines, in which was possible to adjust the temporal discretization. The results showed that using ANN, the accuracy increased 26 times relative to the non-optimized results, also to compare with commercial software the biggest error found was 10%. Therefore, it was proven that when applying the ANN that provided excellent results with lower computational cost.